Hiring!!!

I am looking for self-motivated PhD/Master students to join my L2Opt Group at SMU to conduct research on Learning to Optimize, where deep learning (including LLM) is exploited to solve combinatorial optimization problems, especially the vehicle routing problems. The requirement for SMU PhD admission can be found HERE. Visiting CSC students/researchers, and remote online interns are also welcome. Drop me an email if you are interested!

About Me

I am currently an Assistant Professor at the School of Computing and Information Systems (SCIS), Singapore Management University (SMU). I received the Ph.D from Interdisciplinary Graduate School & School of Computer Science and Engineering, and the M.Sc from School of Electrical and Electronic Engineering, NTU. I received the B.Eng in Automation from Guangdong University of Technology (GDUT), China. I am leading the L2Opt Group at SMU. You may approach me at

Office: Room 5027, 80 Stamford Road, Singapore 178902

E-mail: zgcao [at] smu [dot] edu [dot] sg

Telephone: +65 6826 1341

Recently, my interests focus on Learning to Optimize (L2Opt), also known as Learning for Optimization and closely related to Neural Combinatorial Optimization (NCO), where we exploit Deep (Reinforcement) Learning (including LLM) to solve the classic Combinatorial Optimization Problems (such as Vehicle Routing Problem, Job Shop Scheduling Problem and Bin Packing Problem), Constraint Satisfaction Problem, and Integer Programming Problem. It is a hot topic for both AI and Operations Research, quite interesting yet challenging. Welcome to use our CODE and reach out for collaboration.

Artificial Intelligence: Deep Reinforcement Learning, Deep Learning, Large Language Model (LLM)

Combinatorial Optimization: Routing, Scheduling, Packing; (Mixed) Integer Programming

Applications: Logistics, Transportation, Robotics

News

11.2024: Two papers on L2Opt got accepted at SIGKDD'25.

10.2024: One paper on L2Opt got accepted by TELO.

09.2024: Three papers on L2Opt got accepted at NeurIPS'24.

09.2024: I was titled as the World’s Top 2% Scientist by Stanford University.

08.2024: Our work titled 'Flexible Job Shop Scheduling via Graph Neural Network and Deep Reinforcement Learning' got the Outstanding Paper Award for the IEEE Transactions on Industrial Informatics.

08.2024: I was invited to serve as Area Chair for ICLR'25.

08.2024: One paper on L2Opt got accepted by T-PAMI, where we solved PDP and its variants by integrating Learning to Search (or Improve) and Learning to Construct.

07.2024: I was invited to serve as Area Chair for SIGKDD'25 and Senior PC for AAAI'25.

06.2024: One paper on green VRP got accepted by CAOR, where we solved GVRP with steep roads using improved ALNS.

05.2024: One paper on L2Opt got accepted at SIGKDD'24, where we solved more realistic TSPs using hierarchical neural solvers.

05.2024: Two papers on L2Opt got accepted at ICML'24, where we solved cross-problem vrps using mixture of experts and boosted column generation using adaptive stablization.

04.2024: One paper on L2Opt got accepted at UAI'24, where we improved the neural JSSP solver by learning advanced topological representations.

04.2024: I was invited to serve as Area Chair for NeurIPS'24.

04.2024: One paper on L2Opt got accepted at IJCAI'24, where we proposed cross-problem learning for different VRP variants.

03.2024: Our project - Learning Assisted Human-AI Collaboration for Large-scale Practical Combinatorial Optimization, was awarded by AISG.

03.2024: YI Hang joined the L2Opt Group as a research engineer.

03.2024: One paper on L2Opt got accepted by SMCA, where we proposed dynamic algorithm selection based on reinforcement learning.

02.2024: Two papers on L2Opt got accepted by TNNLS, where we solved VRPs with backhauls and multi-objective VRPs using DRL, respectively.

02.2024: I was invited to serve as Area Chair for SIGKDD'24.

01.2024: I was invited to serve as Area Chair for ICML'24.

01.2024: I gave a talk on Learning to Solve Vehicle Routing Problems to HuaWei (Online).

01.2024: One paper on L2Opt got accepted at ICLR'24, where we solved JSSP using DRL based improvement heuristic.

01.2024: One paper on L2Opt got accepted by T-ASE, where we solved large-scale UAV routing using DRL based divide-and-conquer.

01.2024: One paper on L2Opt got accepted by T-ITS, where we solved stochastic shortest path problem using actor-critic based RL.

12.2023: One paper on L2Opt got accepted at AAMAS'24, where we solved multi-objective VRPs, including MOTSP and MOCVRP.

12.2023: One paper on L2Opt got accepted at AAAI'24, where we solved large-scale VRPs (including TSP, ATSP, CVRP, and PCTSP) in real-time.

10.2023: I gave a talk on Learning to Solve Vehicle Routing Problems to Nanjing University (Online).

09.2023: Six papers on L2Opt got accepted at NeurIPS'23.

09.2023: LIU Yunduo and LIAO Shubing joined the L2Opt Group as visiting students.

07.2023: I joined the School of Computing and Information Systems, SMU, as an Assistant Professor.